Predicting the SYM‐H Index Using the Ring Current Energy Balance Mechanism

Abstract The geomagnetic disturbance index SYM‐H is primarily determined by the total kinetic energy of ring current particles. Therefore, the energy balance mechanism of the ring current can be used to construct an SYM‐H evolution equation for prediction purposes. This study extends a modeling concept developed by Ji et al. (2023), https://doi.org/10.1029/2022ea002560 to establish an algebraic equation for predicting the SYM‐H index based on equilibrium between energy injection and ring current loss. The loss term in the model is determined by a fully connected neural network. The fundamental form of the energy injection function is derived from existing solar wind–magnetosphere energy coupling functions, with its scale factor adjusted as a free‐fitting parameter to optimize the prediction of observations. After being trained on solar wind and SYM‐H observations from 20 magnetic storms, the new model predicts the SYM‐H index well 1 hr and 2 hr in advance, with root mean square errors of 6.7 and 8.9 nT, respectively. These accuracies represent a 7 (1‐hr model) and a 6 (2‐hr model) improvement over the previous model. Furthermore, the scale factors for the solar wind parameters in the energy coupling function determined by the new model can be explained by the previous observations in the magnetic tail current sheet, confirming that the ring current energy primarily originates from the current sheet. The lifetime of the ring current particles, as determined by the neural network, varies with the SYM‐H index. It is approximately 6 hr for the fast recovery phase and more than 10 hr for the slow recovery phase, consistent with the dominant ring current particles changing from oxygen ions to protons during intense storms.